{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:7CDTLE7OKLXMONMIHC7IZI3DVP","short_pith_number":"pith:7CDTLE7O","schema_version":"1.0","canonical_sha256":"f8873593ee52eec7358838be8ca363abd2167c6ab9673fd57f023134e54bc9b9","source":{"kind":"arxiv","id":"2605.23618","version":1},"attestation_state":"computed","paper":{"title":"Benchmarking Google Embeddings 2 against Open-Source Models for Multilingual Dense Retrieval and RAG Systems","license":"http://creativecommons.org/licenses/by-sa/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Domenico Desiato, Giandomenico Solimando, Giuseppe Polese, Stefano Cirillo","submitted_at":"2026-05-22T13:25:13Z","abstract_excerpt":"We benchmark Google Embeddings (GE2), a Vertex-AI-hosted bi-encoder with 2,048-token context and explicit task-type conditioning, against five open-source alternatives: BGE-M3, E5-large, Multilingual-E5-large (mE5-L), LaBSE, and Paraphrase-Multilingual-MPNet (mMPNet). Evaluation covers four BEIR subsets, a synthetic Italian RAG corpus, a chunking ablation considering 5 sizes of tokens with three strategies, and per-query latency on commodity CPU hardware. GE2 ranks first on every task, achieving BEIR avg.nDCG@10 = 0.638 and IT-RAG-Bench nDCG@10 = 0.282, but at 231.6 ms median latency, it is ro"},"verification_status":{"content_addressed":true,"pith_receipt":true,"author_attested":false,"weak_author_claims":0,"strong_author_claims":0,"externally_anchored":false,"storage_verified":false,"citation_signatures":0,"replication_records":0,"graph_snapshot":true,"references_resolved":false,"formal_links_present":false},"canonical_record":{"source":{"id":"2605.23618","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by-sa/4.0/","primary_cat":"cs.CL","submitted_at":"2026-05-22T13:25:13Z","cross_cats_sorted":[],"title_canon_sha256":"b57b8aeba59fd4d68fcf8370c48222dad3d51e20f50e2fe0c4807b7f5c3235ee","abstract_canon_sha256":"d8daf69678e510be983b3779efb3fc9cfda48bebeacce11bd69e26267d480c39"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-25T02:02:22.288657Z","signature_b64":"USuqk/wCOYhJErrYOZMzoYAHmZA9KfQLES6PoI8/3ecQftfEGVKJpjQ2E8c1W4DkLfkMyAuFnYeVFJ3Ei1xlBw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"f8873593ee52eec7358838be8ca363abd2167c6ab9673fd57f023134e54bc9b9","last_reissued_at":"2026-05-25T02:02:22.287918Z","signature_status":"signed_v1","first_computed_at":"2026-05-25T02:02:22.287918Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Benchmarking Google Embeddings 2 against Open-Source Models for Multilingual Dense Retrieval and RAG Systems","license":"http://creativecommons.org/licenses/by-sa/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Domenico Desiato, Giandomenico Solimando, Giuseppe Polese, Stefano Cirillo","submitted_at":"2026-05-22T13:25:13Z","abstract_excerpt":"We benchmark Google Embeddings (GE2), a Vertex-AI-hosted bi-encoder with 2,048-token context and explicit task-type conditioning, against five open-source alternatives: BGE-M3, E5-large, Multilingual-E5-large (mE5-L), LaBSE, and Paraphrase-Multilingual-MPNet (mMPNet). Evaluation covers four BEIR subsets, a synthetic Italian RAG corpus, a chunking ablation considering 5 sizes of tokens with three strategies, and per-query latency on commodity CPU hardware. GE2 ranks first on every task, achieving BEIR avg.nDCG@10 = 0.638 and IT-RAG-Bench nDCG@10 = 0.282, but at 231.6 ms median latency, it is ro"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.23618","kind":"arxiv","version":1},"verdict":{"id":null,"model_set":{},"created_at":null,"strongest_claim":"","one_line_summary":"","pipeline_version":null,"weakest_assumption":"","pith_extraction_headline":""},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2605.23618/integrity.json","findings":[],"available":true,"detectors_run":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938"},"references":{"count":0,"sample":[],"resolved_work":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","internal_anchors":0},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"},"aliases":[{"alias_kind":"arxiv","alias_value":"2605.23618","created_at":"2026-05-25T02:02:22.288036+00:00"},{"alias_kind":"arxiv_version","alias_value":"2605.23618v1","created_at":"2026-05-25T02:02:22.288036+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.23618","created_at":"2026-05-25T02:02:22.288036+00:00"},{"alias_kind":"pith_short_12","alias_value":"7CDTLE7OKLXM","created_at":"2026-05-25T02:02:22.288036+00:00"},{"alias_kind":"pith_short_16","alias_value":"7CDTLE7OKLXMONMI","created_at":"2026-05-25T02:02:22.288036+00:00"},{"alias_kind":"pith_short_8","alias_value":"7CDTLE7O","created_at":"2026-05-25T02:02:22.288036+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/7CDTLE7OKLXMONMIHC7IZI3DVP","json":"https://pith.science/pith/7CDTLE7OKLXMONMIHC7IZI3DVP.json","graph_json":"https://pith.science/api/pith-number/7CDTLE7OKLXMONMIHC7IZI3DVP/graph.json","events_json":"https://pith.science/api/pith-number/7CDTLE7OKLXMONMIHC7IZI3DVP/events.json","paper":"https://pith.science/paper/7CDTLE7O"},"agent_actions":{"view_html":"https://pith.science/pith/7CDTLE7OKLXMONMIHC7IZI3DVP","download_json":"https://pith.science/pith/7CDTLE7OKLXMONMIHC7IZI3DVP.json","view_paper":"https://pith.science/paper/7CDTLE7O","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2605.23618&json=true","fetch_graph":"https://pith.science/api/pith-number/7CDTLE7OKLXMONMIHC7IZI3DVP/graph.json","fetch_events":"https://pith.science/api/pith-number/7CDTLE7OKLXMONMIHC7IZI3DVP/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/7CDTLE7OKLXMONMIHC7IZI3DVP/action/timestamp_anchor","attest_storage":"https://pith.science/pith/7CDTLE7OKLXMONMIHC7IZI3DVP/action/storage_attestation","attest_author":"https://pith.science/pith/7CDTLE7OKLXMONMIHC7IZI3DVP/action/author_attestation","sign_citation":"https://pith.science/pith/7CDTLE7OKLXMONMIHC7IZI3DVP/action/citation_signature","submit_replication":"https://pith.science/pith/7CDTLE7OKLXMONMIHC7IZI3DVP/action/replication_record"}},"created_at":"2026-05-25T02:02:22.288036+00:00","updated_at":"2026-05-25T02:02:22.288036+00:00"}